Re: better algorithm: bitwise reordering

I'm sorry, but I'm unable to understand this post.

Could you please specify exactly what transformations you have to work with, and correct your diagrams so that the transformations pictured are the ones that happened? For instance where you say, swap rows 0 and 1 I'm having no luck understanding what transformation is supposed to have happened.

Comment on Re: better algorithm: bitwise reordering

Replies are listed 'Best First'.
Re^2: better algorithm: bitwise reordering by hv (Prior) on Aug 25, 2004 at 16:41 UTC
Apologies for the lack of clarity. I neglected to mention that after swapping the rows I also reordered the columns so that they were in ascending order of the bitwise representation of their stars. So before: `. . . x * * * (row 2, bit value 4) . . . * * * * (row 1, bit value 2) . * * . . . * (row 0, bit value 1) 0 1 1 2 6 6 7 (bitwise column values)` [download] after swapping rows 0 and 1: `. . . x * * * (still row 2) . * * . . . * (was row 0, now row 1) . . . * * * * (was row 1, now row 0) 0 2 2 1 5 5 7 (new bitwise column values)` [download] and after reordering columns: `. x . . * * * . . * * . . * . * . . * * * 0 1 2 2 5 5 7 (new bitwise column values)` [download] Similarly the second swap gives columns (0, 1, 4, 4, 3, 3, 7), and reordering the columns gives (0, 1, 3, 3, 4, 4, 7). Does that make it any clearer? Hugo	[reply] [d/l] [select]
Re^3: better algorithm: bitwise reordering by tilly (Archbishop) on Aug 25, 2004 at 20:06 UTC
That makes it a lot clearer - for a start I realize that you're counting rows from the bottom up, not the top down!	[reply]